Nat has chaired the O'Reilly Open Source Convention and other O'Reilly conferences for over a decade. He ran the first web server in New Zealand, co-wrote the best-selling Perl Cookbook, and was one of the founding Radar bloggers. He lives in New Zealand and consults in the Asia-Pacific region.
The Sad State of Sysadmin in the Age of Containers (Erich Schubert) — a Grumpy Old Man rant, but solid. And since nobody is still able to compile things from scratch, everybody just downloads precompiled binaries from random websites. Often without any authentication or signature.
Pinball — Pinterest open-sourced their data workflow manager.
Disambiguating Databases (ACM) — The scope of the term database is vast. Technically speaking, anything that stores data for later retrieval is a database. Even by that broad definition, there is functionality that is common to most databases. This article enumerates those features at a high level. The intent is to provide readers with a toolset with which they might evaluate databases on their relative merits.
Hello Barbie — I just can’t imagine a business not wanting to mine and repurpose the streams of audio data coming into their servers. “You listen to Katy Perry a lot. So do I! You have a birthday coming up. Have you told your parents about the Katy Perry brand official action figurines from Mattel? Kids love ’em, and demo data and representative testing indicates you will, too!” Or just offer a subscription service where parents can listen in on what their kids say when they play in the other room with their friends. Or identify product mentions and cross-market offline. Or …
There Is No Now — One of the most important results in the theory of distributed systems is an impossibility result, showing one of the limits of the ability to build systems that work in a world where things can fail. This is generally referred to as the FLP result, named for its authors, Fischer, Lynch, and Paterson. Their work, which won the 2001 Dijkstra Prize for the most influential paper in distributed computing, showed conclusively that some computational problems that are achievable in a “synchronous” model in which hosts have identical or shared clocks are impossible under a weaker, asynchronous system model.
Deep Learning Hardware Guide — One of the worst things you can do when building a deep learning system is to waste money on hardware that is unnecessary. Here I will guide you step by step through the hardware you will need for a cheap high performance system.
As a Working Manager (Ian Bicking) — I look forward to every new entry in Ian’s diary, and this one didn’t disappoint. But I’m a working manager. Is now the right time to investigate that odd log message I’m seeing, or to think about who I should talk to about product opportunities? There’s no metric to compare the priority of two tasks that are so far apart. If I am going to find time to do development I am a bit worried I have two options: (1) Keep doing programming after hours; (2) Start dropping some balls as a manager.
Introducing Yosemite (Facebook) — a modular chassis that contains high-powered system-on-a-chip (SoC) processor cards.
The Joyless World of Data-Driven Startups — There is so much invisible, fluid context wrapped around a data point that we are usually unable to fully comprehend exactly what that data represents or means. We often think we know, but we rarely do. But we really WANT it to mean something, because using data in our work is scientific. It’s not our decision that was wrong — we used the data that was available. Data is the ultimate scapegoat.
History of the Urban Dashboard — the dashboard and its user had to evolve in response to one another. The increasing complexity of the flight dashboard necessitated advanced training for pilots — particularly through new flight simulators — and new research on cockpit design.
Surgical Micro-Robot Swarms — A swarm of medical microrobots. Start with cm sized robots. These already exist in the form of pillbots and I reference the work of Paolo Dario’s lab in this direction. Then get 10 times smaller to mm sized robots. Here we’re at the limit of making robots with conventional mechatronics. The almost successful I-SWARM project prototyped remarkable robots measuring 4 x 4 x 3mm. But now shrink by another 3 orders of magnitude to microbots, measured in micrometers. This is how small robots would have to be in order to swim through and access (most of) the vascular system. Here we are far beyond conventional materials and electronics, but amazingly work is going on to control bacteria. In the example I give from the lab of Sylvain Martel, swarms of magnetotactic bacteria are steered by an external magnetic field and, interestingly, tracked in an MRI scanner.
Media Hacking — interesting discussion of the techniques used to spread disinformation through social media, often using bots to surface/promote a message.
Apple Research Kit — Apple positioning their mobile personal biodata tools with medical legitimacy, presumably as a way to distance themselves from the stereotypical quantified selfer. I’m reminded of the gym chain owner who told me, about the Nike+, “yeah, maybe 5% of my clients will want this. The rest go to the gym so they can eat and drink what they want.”
Designing the Human-Robot Relationship (O’Reilly) — We can use those same principles [Jakob Nielsen’s usability heuristics] and look for implications of robots serving our higher ordered needs, as we move from serving needs related to convenience or performance to actually supporting our decision making to emerging technologies, moving from being able to do anything or be magic in terms of the user interface to being more human in the user interface.
Why Are Geospatial Databases So Hard To Build? — Algorithms in computer science, with rare exception, leverage properties unique to one-dimensional scalar data models. In other words, data types you can abstractly represent as an integer. Even when scalar data types are multidimensional, they can often be mapped to one dimension. This works well, as the majority of [what] data people care about can be represented with scalar types. If your data model is inherently non-scalar, you enter an algorithm wasteland in the computer science literature.
The Web’s Grain (Frank Chimero) — What would happen if we stopped treating the web like a blank canvas to paint on, and instead like a material to build with?
Bruce Sterling on Convergence of Humans and Machines — I like to use the terms “cognition” and “computation”. Cognition is something that happens in brains, physical, biological brains. Computation is a thing that happens with software strings on electronic tracks that are inscribed out of silicon and put on fibre board. They are not the same thing, and saying that makes the same mistake as in earlier times, when people said that human thought was like a steam engine.
Smart Pocket Watch — I love to see people trying different design experiences. This is beautiful. And built on Firefox OS!
Knowledge-Based Trust (PDF) — Google research paper on how to assess factual accuracy of web page content. It was bad enough when Google incentivised people to make content-free pages. Next there’ll be a reward for scamming bogus facts into Google’s facts database.
Microservices in Go — tale of rewriting a Ruby monolith as Go microservices. Interesting, though being delivered at Gophercon India suggests the ending is probably not unhappy.
Watch & Wear (John Cross Neumann) — Android watch as predictor of the value and experience of an Apple Watch. I believe this is the true sweet spot for meaningful wearable experiences. Information that matters to you in the moment, but requires no intervention. Wear actually does this extremely well through Google Now. Traffic, Time to Home, Reminders, Friend’s Birthdays, and Travel Information all work beautifully. […] After some real experience with Wear, I think what is more important is to consider what Apple Watch is missing: Google Services. Google Services are a big component of what can make wearing a tiny screen on your wrist meaningful and personal. I wouldn’t be surprised after the initial wave of apps through the app store if Google Now ends up being the killer app for Apple Watch.
Solving 11 Likely Problems In Your Multithreaded Code (Joe Duffy) — a good breakdown of concurrency problems, including lower-level ones than high-level languages expose. But beware. If you try this [accessing variables with synchronisation] on a misaligned memory location, or a location that isn’t naturally sized, you can encounter a read or write tearing. Tearing occurs because reading or writing such locations actually involves multiple physical memory operations. Concurrent updates can happen in between these, potentially causing the resultant value to be some blend of the before and after values.
Obama Sharply Criticizes China’s Plans for New Technology Rules (Reuters) — In an interview with Reuters, Obama said he was concerned about Beijing’s plans for a far-reaching counterterrorism law that would require technology firms to hand over encryption keys, the passcodes that help protect data, and install security “backdoors” in their systems to give Chinese authorities surveillance access. Goose sauce is NOT gander sauce! NOT! Mmm, delicious spook sauce.
You Guys Realize the Apple Watch is Going to Flop, Right? — leaving aside the “guys” assumption of its readers, you can take this either as a list of the challenges Apple will inevitably overcome or bypass when they release their watch, or (as intended) a list of the many reasons that it’s too damn soon for watches to be useful. The Apple Watch is Jonathan Ive’s new Newton. It’s a potentially promising form that’s being built about 10 years before Apple has the technology or infrastructure to pull it off in a meaningful way. As a result, the novel interactions that could have made the Apple watch a must-have device aren’t in the company’s launch product, nor are they on the immediate horizon. And all Apple can sell the public on is a few tweets and emails on their wrists—an attempt at a fashion statement that needs to be charged once or more a day.
InfluxDB, Now With Tags and More Unicorns — The combination of these new features [tagging, and the use of tags in queries] makes InfluxDB not just a time series database, but also a database for time series discovery. It’s our solution for making the problem of dealing with hundreds of thousands or millions of time series tractable.
The End of Apps as We Know Them — It may be very likely that the primary interface for interacting with apps will not be the app itself. The app is primarily a publishing tool. The number one way people use your app is through this notification layer, or aggregated card stream. Not by opening the app itself. To which one grumpy O’Reilly editor replied, “cards are the new walled garden.”
Signal 2.0 — Signal uses your existing phone number and address book. There are no separate logins, usernames, passwords, or PINs to manage or lose. We cannot hear your conversations or see your messages, and no one else can either. Everything in Signal is always end-to-end encrypted, and painstakingly engineered in order to keep your communication safe.
Machine Learning Done Wrong — When dealing with small amounts of data, it’s reasonable to try as many algorithms as possible and to pick the best one since the cost of experimentation is low. But as we hit “big data,” it pays off to analyze the data upfront and then design the modeling pipeline (pre-processing, modeling, optimization algorithm, evaluation, productionization) accordingly.
Ten Simple Rules for Lifelong Learning According to Richard Hamming (PLoScompBio) — Exponential growth of the amount of knowledge is a central feature of the modern era. As Hamming points out, since the time of Isaac Newton (1642/3-1726/7), the total amount of knowledge (including but not limited to technical fields) has doubled about every 17 years. At the same time, the half-life of technical knowledge has been estimated to be about 15 years. If the total amount of knowledge available today is x, then in 15 years the total amount of knowledge can be expected to be nearly 2x, while the amount of knowledge that has become obsolete will be about 0.5x. This means that the total amount of knowledge thought to be valid has increased from x to nearly 1.5x. Taken together, this means that if your daughter or son was born when you were 34 years old, the amount of knowledge she or he will be faced with on entering university at age 17 will be more than twice the amount you faced when you started college.